About the trimming of poor quality bases from fastq files during the usage of mothur

Hi there,

I am very glad to use mothur to analyse my 16s rRNA data but after reading through the MiSeq SOP on the following link:


I have one question about the quality trimming.

I couldn’t find any step that is dedicated to trim the fastq files before merging the R1 and R2 reads based on the base quality of fastq files.

Does it mean that there is no need to trim any poor quality bases of fastq files during the whole process? Why?

Many thanks for your help,

Kind regards


This should really depend on the data that you have. Consider the length of your reads and inserts and the overall quality of your sequencing data.
If you have average quality data and at least some overlap of your reads, e.g. when sequencing V3V4 and 2x300 cycles on MiSeq, you should be fine in most cases. The quality drop at the 3’ end is (at least partially) compensated by the overlap.

That’s why it is recommend to use protocols that maximize the overlap instead of throwing away lower quality data at the read ends. See this blog entry for example: http://blog.mothur.org/2014/09/11/Why-such-a-large-distance-matrix/

Thanks for chiming in Andreas - indeed - we don’t recommend trimming the reads before assembly. In fact, doing so makes the error rates much higher.


This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.